Generally, when performing image recognition, detection of a detection target is performed by collating a feature value of an image extracted from a captured image with a feature value of a registration image registered in a dictionary in advance.